<?xml version="1.0" encoding="UTF-8"?>
<merge-requests type="array">
  <merge-request>
    <created-at type="datetime">2010-02-09T12:17:00Z</created-at>
    <ending-commit>06b631f435bf6184cec13fddcce8ea4ac453a8b1</ending-commit>
    <id type="integer">2992</id>
    <proposal>When showing the poup its minimumWidth is now taken into account.

If the popup would not fit on screen, the x position is adjusted accordingly.</proposal>
    <summary>Make QCompleter respect the minimum width of the popup widget when showing it (QTBUG-2726)</summary>
    <updated-at type="datetime">2010-02-09T12:17:02Z</updated-at>
    <status>New</status>
    <username>~njeisecke</username>
    <source_repository>
      <name>njeiseckes-clone</name>
      <branch>njeiseck/qcompleter_minimum_size_mr_2</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-02-08T07:31:57Z</created-at>
    <ending-commit>9c71a3e7221c0c1607e07ba1a437a5a35c6a2863</ending-commit>
    <id type="integer">2970</id>
    <proposal>A QGraphicsItem is using setCacheMode(QGraphicsItem::ItemCoordinateCache). When item invalidates only partial region, its updated region gets painted on top of the old content. If the item is non-opaque, the old content is shown below the newly painted content.

Patch contains one-line fix and a unittest</proposal>
    <summary>Fix partial updates of non-opque QGraphicsItems [QTBUG-7863]</summary>
    <updated-at type="datetime">2010-02-08T07:32:17Z</updated-at>
    <status>New</status>
    <username>~kimmok</username>
    <source_repository>
      <name>kimmoks-qt</name>
      <branch>fix-bug-7863-cached-item-partial-update</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-02-04T07:34:25Z</created-at>
    <ending-commit>c2652a8aa1c61017bd30eaf7bffa0c92f2e75dab</ending-commit>
    <id type="integer">2929</id>
    <proposal>the following code shows the issue: 
//------------------------------------------------------------------------------------------------
QLocale locale = QLocale(QLocale::English);
//QString dateTimeStr = QString("Monday, January 5, 2009 11:48:32 AM"); // with no blank space char after "AM", this date time str cannot be parsed correctly
QString dateTimeStr = QString("Monday, January 5, 2009 11:48:32 AM ");

//QDateTime value = locale.toDateTime(dateTimeStr , locale.dateTimeFormat(QLocale::LongFormat)); // this will success
QDateTime value = locale.toDateTime(dateTimeStr , QLocale::LongFormat); // cannot return right QDateTime value
//------------------------------------------------------------------------------------------------
souce code of "toDateTime"
QDateTime QLocale::toDateTime(const QString &amp;string, FormatType format) const

{ return toDateTime(string, dateFormat(format)); // "dateFormat" should be replaced by "dateTimeFormat" }
</proposal>
    <summary>QLocale::toDateTime cannot convert value correctly</summary>
    <updated-at type="datetime">2010-02-09T10:13:09Z</updated-at>
    <status>New</status>
    <username>~suuuushe</username>
    <source_repository>
      <name>bc-qt</name>
      <branch>QTBUG-7898</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-29T18:24:43Z</created-at>
    <ending-commit>a7af14c8cd922ab9db5e67e725128b2c0f5c9d25</ending-commit>
    <id type="integer">2852</id>
    <proposal>Fixed bug QTBUG-5448 by not only invalidating the parent widget's top layout, but also the layout hierarchy between the changed widget and the parent widget.

There's another merge request - #2296 - that one can be removed.</proposal>
    <summary>Fixed QTBUG-5448, problems with nested layouts - second try</summary>
    <updated-at type="datetime">2010-02-05T15:13:26Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~cschleifenbaum</username>
    <source_repository>
      <name>kdab-for-upstream</name>
      <branch>nested-layouts-fix</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-28T03:20:41Z</created-at>
    <ending-commit>48142dcc74b26a952a3fcb4db9421ab3075b1a14</ending-commit>
    <id type="integer">2829</id>
    <proposal>Modified _scheme() and parse() methods to handle wrong schemes. Method _scheme() now returns a bool that is true if the scheme is valid, and false otherwise. The boolean variable is false only if the url starts with anything other than an alpha character AND is the type of url with ':' (used in http, etc.).</proposal>
    <summary>Patch to QTBUG-3168</summary>
    <updated-at type="datetime">2010-02-02T11:05:10Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~carolinagomes</username>
    <source_repository>
      <name>carolinagomes-qt</name>
      <branch>patch-branch</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-26T14:25:54Z</created-at>
    <ending-commit>a0b69debb32b0528d6423f28ac4695c5bdab1ac8</ending-commit>
    <id type="integer">2811</id>
    <proposal>If a client connects with an invalid certificate that doesn't parse properly, this will generate errors that are reported only through the thread error queue. Unless this queue is drained, the next call to SSL_get_error(), which very likely is from a different QSslSocket, will report these additional errors.

Thus, a malicious client connecting to a central server can cause other clients to be disconnected.

Please note that this is a hack; a stopgap measure until a proper solution can be found.</proposal>
    <summary>Avoid SSL error poisoning</summary>
    <updated-at type="datetime">2010-01-26T14:25:55Z</updated-at>
    <status>New</status>
    <username>~slicer</username>
    <source_repository>
      <name>slicers-clone</name>
      <branch>nosslpoison</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-23T15:30:34Z</created-at>
    <ending-commit>2af22a00b28c0d1e092a7e6b03e529b2ee4d2f3e</ending-commit>
    <id type="integer">2779</id>
    <proposal>Improvements to internal QZipReader class

* Add ability to read last modification time of zip entry 
* Add isValid() method to QZipReader::FileInfo 
* Add device() method to QZip* classes</proposal>
    <summary>Improvements to internal QZipReader class</summary>
    <updated-at type="datetime">2010-01-23T15:39:37Z</updated-at>
    <status>Reviewing</status>
    <username>~ritt.k</username>
    <source_repository>
      <name>ritts_qt</name>
      <branch>improvements/qzip</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-21T01:15:37Z</created-at>
    <ending-commit>2d6b16b9dd565537469910514dd9e9007c1b8cee</ending-commit>
    <id type="integer">2752</id>
    <proposal>Patch to bug in QtDBus - "writeProxy() not generating signals correctly in all cases". Modified writeProxy(), writeAdaptor(), makeArgName and writeArgList() of qdbusxml2cpp.cpp, in order to allow the correct "out" parameters processing for signals.</proposal>
    <summary>Patch to QTBUG-914</summary>
    <updated-at type="datetime">2010-01-27T14:24:10Z</updated-at>
    <status>New</status>
    <username>~carolinagomes</username>
    <source_repository>
      <name>carolinagomes-qt</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-20T12:22:11Z</created-at>
    <ending-commit>ac0850d196061d782aea9788d7d0faede9a3562c</ending-commit>
    <id type="integer">2740</id>
    <proposal>The dependency for now was fully removed and qtconfig tool now successfully build with qt3support turned off.

A have not targeted for complete refactoring of the tool, but in a process of work I've removed some legacy and unused code and done some other small modifications (see the comments on the commits).</proposal>
    <summary>Removed the qtconfig's dependency from qt3support</summary>
    <updated-at type="datetime">2010-02-05T16:34:22Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~bobs</username>
    <source_repository>
      <name>qtconfig-refactor</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-19T10:35:59Z</created-at>
    <ending-commit>0a4ea17646448af50bf9140c24207114616978fd</ending-commit>
    <id type="integer">2725</id>
    <proposal>Related to: QTBUG-7415

Since qdbusviewer only matches methods by name, overloaded methods can not be called. This patch adds D-bus type signature matching in addition to the name match.</proposal>
    <summary>Type signature based method identification for qdbusviewer</summary>
    <updated-at type="datetime">2010-01-19T10:36:00Z</updated-at>
    <status>New</status>
    <username>~kko</username>
    <source_repository>
      <name>kkos-qt</name>
      <branch>qdbusviewer-signaturecheck</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-15T10:17:34Z</created-at>
    <ending-commit>eecf29665e52958ae9ba97838ee26844c8b54b8a</ending-commit>
    <id type="integer">2681</id>
    <proposal>I added some additional checks to finding a method in qdbus command line tool. Previously, it would take the first method matching the name, so in case of

void foo(int a)
void foo(int a, QStringList b)

and calling

qdbus org.dbustest / foo 15 "(" string list ")"

It would result in "invalid number of parameters" since it tried to match against the first method.</proposal>
    <summary>Qdbus method overloading support</summary>
    <updated-at type="datetime">2010-01-15T10:17:41Z</updated-at>
    <status>New</status>
    <username>~kko</username>
    <source_repository>
      <name>kkos-qt</name>
      <branch>dbus-overloading</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-08T21:56:43Z</created-at>
    <ending-commit>d2fb9848f6007776477773a0946541f5af2d1108</ending-commit>
    <id type="integer">2607</id>
    <proposal>this also removes dirty trick from QFileInfo</proposal>
    <summary>don't use owner's permissions for current user; calculate them instead</summary>
    <updated-at type="datetime">2010-01-29T10:01:20Z</updated-at>
    <status>New</status>
    <username>~ritt.k</username>
    <source_repository>
      <name>ritts_qt</name>
      <branch>labs/qfsfileengine-calculate_user_permissions</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-08T17:15:48Z</created-at>
    <ending-commit>95152cb6d31cacfbe94d13b14af82014fd2abf8e</ending-commit>
    <id type="integer">2603</id>
    <proposal></proposal>
    <summary>minor fixes and clean-ups for QSyntaxHighlighter</summary>
    <updated-at type="datetime">2010-01-12T14:45:07Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~ritt.k</username>
    <source_repository>
      <name>ritts_qt</name>
      <branch>fixes/qsyntaxhighlighter/minor</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-08T16:20:10Z</created-at>
    <ending-commit>c2acf4b49ec197af586fac4b02bf4f4692a7a1b8</ending-commit>
    <id type="integer">2601</id>
    <proposal>plus few minor fixes</proposal>
    <summary>improve high unicodes support in QTextEngine</summary>
    <updated-at type="datetime">2010-01-08T16:20:12Z</updated-at>
    <status>New</status>
    <username>~ritt.k</username>
    <source_repository>
      <name>ritts_qt</name>
      <branch>fixes/qunicode/qtextengine-high_codes_sipport</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2010-01-07T23:22:58Z</created-at>
    <ending-commit>987813649da3e4c5e7f135220d50bd071a47e063</ending-commit>
    <id type="integer">2594</id>
    <proposal>The current formula for isLeapYear() assumes that years -4, -8, etc are
leap years in the Julian Calendar. This is not the case, leap years in
fact fall in -1, -5, -9, etc as there is no year 0 in the Julian
Calendar.</proposal>
    <summary>Fix QDate::isLeapYear() for years &lt; 1</summary>
    <updated-at type="datetime">2010-01-11T19:35:21Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~odysseus</username>
    <source_repository>
      <name>odysseus-clone</name>
      <branch>qdatetime-fixes</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-30T15:24:17Z</created-at>
    <ending-commit>7d02d34f87e8388401e2cae82f0ab828500aa67b</ending-commit>
    <id type="integer">2528</id>
    <proposal>Take advantage of the IDropTargetHelper interface as described in:
http://msdn.microsoft.com/en-us/library/ms997500.aspx

This would be really appreciated especially on Windows Vista/7, which provides really nice file &amp; folder images while dragging from Windows Explorer</proposal>
    <summary>Add support for Shell style drag 'n drop on Windows (IDropTargetHelper)</summary>
    <updated-at type="datetime">2009-12-30T20:38:06Z</updated-at>
    <status>New</status>
    <username>~miniak</username>
    <source_repository>
      <name>miniak</name>
      <branch>IDropTargetHelper</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-27T17:49:43Z</created-at>
    <ending-commit>65d36199214dfd180a0ca557e64122cdc06da07a</ending-commit>
    <id type="integer">2503</id>
    <proposal>This patch adds support for 8Track e-ink displays in the QWS framebuffer driver.</proposal>
    <summary>Support 8Track e-Ink devices in the framebuffer driver</summary>
    <updated-at type="datetime">2010-01-11T11:40:06Z</updated-at>
    <status>Reviewing</status>
    <username>~blankpage</username>
    <source_repository>
      <name>blankpage-patches</name>
      <branch>eink</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-24T01:33:56Z</created-at>
    <ending-commit>06781f7ff63f2af7aacaa9028da513fecc81ecc0</ending-commit>
    <id type="integer">2488</id>
    <proposal>This series fixes the build on OpenSolaris, using the compiler available in the Sun Studio package.

Patch 1: OpenGL is an issue on these systems, as the library internally uses symbols from the `Xdamage' library.  Thus applications that use OpenGL must also link in Xdamage, even if the app does not use the library itself.

Patch 2: Related to patch 1.  To get the OpenGL configuration test program to work, we need Xdamage.  Otherwise the link will fail and the configure script will incorrectly think OpenGL is broken on the system.  Use QMAKE_LIBS_OPENGL, instead of hardcoding -lGL, which will pick up Xdamage because of our previous commit.

Patch 3: These systems come with both Sun's compiler and gcc.  "cc" actually ends up referring to gcc.  This is not a problem with the C++ compiler, because gcc does not install a symlink/frontend with the name "CC".  To fix this, explicitly set the path to Sun's "cc", ensuring we use that C compiler.  Without this patch, we end up passing e.g. -KPIC to the GNU compiler, which errors out.

qtriangulator.cpp also fails to compile, but in writing a summary for the patch I have, I noticed that my fix was incorrect.  Will send another merge request for it.</proposal>
    <summary>Build fixes for Sun CC on OpenSolaris</summary>
    <updated-at type="datetime">2010-01-11T14:37:17Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~tfogal</username>
    <source_repository>
      <name>tfogal-qt</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-22T23:48:28Z</created-at>
    <ending-commit>cd83ba28722cfa528c8c0b5d20a9ca072091e176</ending-commit>
    <id type="integer">2470</id>
    <proposal>The changes fix task tracker bug 182246 and #QTBUG-6996.
The fix to bug 182246 only fixed the sample programm attached to the bug. It was not a general fix. 

Currently it is not possible to print accurate documents with Qt because of the incorrect rounding performed in qt_real_to_string and other problems in the function. The changes fix the issues and let Qt write PDF files with the highest possible accuracy depending on the machine.

This is resubmitted request 2251 with updated changes to print all qreal numbers as PDF reals.</proposal>
    <summary>Fix task 182246 and bug #QTBUG-6996</summary>
    <updated-at type="datetime">2010-01-23T15:01:03Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~christophbartoschek</username>
    <source_repository>
      <name>christophbartoschek-qt</name>
      <branch>task-182246-5</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-22T03:09:10Z</created-at>
    <ending-commit>502431c7a5043c47da8ebb90042c849116f9cc57</ending-commit>
    <id type="integer">2467</id>
    <proposal>Calculator from folder examples/widgets/calculator was added keyboard input capability (math operations can be typed in now, not just limited to selection of digits and operators by clicking with mouse).

Besides that, a new folder testcalc was added to examples/qtestlib to provide a real working approach to keyboard-driven testing using QTestLib

Please run tests this way:
./testcalc -keydelay 200</proposal>
    <summary>Provide keyboard input to calculator example and add qtestlib example for calculator </summary>
    <updated-at type="datetime">2010-01-06T03:44:34Z</updated-at>
    <status>Reviewing</status>
    <username>~pablojr</username>
    <source_repository>
      <name>pablojr-qt</name>
      <branch>testcalc</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-18T18:40:00Z</created-at>
    <ending-commit>22c4b0ecbd46083cf5fb1f727b9b9ab4f3e4f200</ending-commit>
    <id type="integer">2445</id>
    <proposal>Unrolled loops and made functions to be inline. This gives about 30% on our tests in such operations like drawing text and vertical lines. </proposal>
    <summary>Raster engine performance optimizations. </summary>
    <updated-at type="datetime">2010-01-15T11:09:30Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~amolokoedov</username>
    <source_repository>
      <name>amolokoedov-qt</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-18T14:22:41Z</created-at>
    <ending-commit>588490fbbeca58ef1acbda4946e238a8f834bf57</ending-commit>
    <id type="integer">2441</id>
    <proposal>This set of changes is to enable the capability of setting a fragment's visibility. If a fragment is set to invisible the fragment is skipped during the layout in QTextEngine.

This feature has been added due to the need we felt during the implementation of Change Tracking in KWord.

Three new API's has been added
- setVisible and isVisible to QTextFragment
  These API's set and test the visibility status of a QTextFragment
- visibleText to QTextBlock
  This API returns the visible text of the block i.e only the text that would be  considered for layout.

Changes have been made in QTextEngine to skip the invisible fragments during layout.

Note: Fragment's visibility status only affects the Layout. So a find on a QTextDocument would return a cursor position *without* skipping invisible fragments.

Tests have been added to QTextLayout and QTextBlock to test and verify this functionality</proposal>
    <summary>Fragment Visibility Support</summary>
    <updated-at type="datetime">2009-12-18T14:23:31Z</updated-at>
    <status>New</status>
    <username>~ganeshp</username>
    <source_repository>
      <name>ganesh-qt</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-17T14:01:52Z</created-at>
    <ending-commit>c68cfcee1ae17f75845550850008dab956432904</ending-commit>
    <id type="integer">2435</id>
    <proposal>The patch allows external styles to see the index of the tab they're supposed to draw, to give the style some more possibilities to customize
the look of tab bars (e.g. alternating colors, tabs getting bigger the closer they are to the active tab, ...)
</proposal>
    <summary>Export the tab index to the StyleOptionTab</summary>
    <updated-at type="datetime">2009-12-27T17:43:06Z</updated-at>
    <status>New</status>
    <username>~blankpage</username>
    <source_repository>
      <name>blankpage-patches</name>
      <branch>show_tab_index_to_styleoption</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-17T08:36:13Z</created-at>
    <ending-commit>de259f5f61c0de85c7fe1412b470176e1a31b24e</ending-commit>
    <id type="integer">2432</id>
    <proposal>Besides two minor performance improvement commits, two new features were added:

1) To access the Windows registry, short names for the root keys may be used (HKLM, HKCU etc.) just as in the native Windows API.

2) QSettings was extended to allow accessing the 32-bit registry from a 64-bit application and vice versa (see QTBUG-3845).</proposal>
    <summary>Improvements for QSettings' Windows registry access (fixes QTBUG-3845)</summary>
    <updated-at type="datetime">2010-02-06T12:14:41Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~sschuberth</username>
    <source_repository>
      <name>sschuberths-clone</name>
      <branch>qsettings-registry-32-64-bit</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-16T00:56:40Z</created-at>
    <ending-commit>2780bcd15bdf4ac22ce2a49ecb8ca6cc71e0cd95</ending-commit>
    <id type="integer">2420</id>
    <proposal>Apply patch http://svn.clazzes.org/svn/mingw-pkg/trunk/mingw-w64-deb/mingw-w64-qt4-4.5.3/patches/qt-4.5.2-mingw-w64-fix-hwnd-64bit-casts.patch

This fixes some 64bit cast issues that are picked up by gcc 4.5
on windows 64bit.</proposal>
    <summary>Fixed hwnd 64bit casts</summary>
    <updated-at type="datetime">2009-12-16T00:56:53Z</updated-at>
    <status>New</status>
    <username>~256fx</username>
    <source_repository>
      <name>256fxs-qt</name>
      <branch>4.5-mingw-w64</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.5</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-14T00:03:50Z</created-at>
    <ending-commit>eca5eadfbb5b21c673936793cf081536316db7ed</ending-commit>
    <id type="integer">2395</id>
    <proposal>QT_LFLAGS_PSQL can introduce -lz. To respect qt-zlib, this must be
be removed so that  zlib will come from -lQtCore4. Otherwise, linking
can fail due to there being two different libraries providing zlib.
This affects building shared Qt.

This patch fixes the problem for win32-g++ and also unix. Assumed
that QT_LFLAGS_PSQL is set in configure step if -lpq is not enough.

Not sure why parallel logic is in src/sql/drivers/drivers.pri but
I fixed that too.</proposal>
    <summary>remove -lz from QT_LFLAGS_PSQL so as to respect qt-zlib</summary>
    <updated-at type="datetime">2010-01-11T15:03:45Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~mabrand</username>
    <source_repository>
      <name>mabrands-clone</name>
      <branch>fixes</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-13T23:59:52Z</created-at>
    <ending-commit>7228fdef61fd126481574e53d6d022d134219cc3</ending-commit>
    <id type="integer">2394</id>
    <proposal>OPENSSL_LIBS can introduce -lz. To respect qt-zlib, this must be
replaced with -lQtCore4. Otherwise, linking can fail due to there
being two different libraries providing zlib.</proposal>
    <summary>remove -lz from OPENSSL_LIBS so as to respect qt-zlib</summary>
    <updated-at type="datetime">2010-01-11T15:07:08Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~mabrand</username>
    <source_repository>
      <name>mabrands-clone</name>
      <branch>fixes</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-11T16:20:42Z</created-at>
    <ending-commit>282155a67bc012bad2d2ab617179733e8694950e</ending-commit>
    <id type="integer">2374</id>
    <proposal>    QHttpNetworkReplyPrivate::readBody was always reading as much as possible
    from the underlying socket, so even setting the buffer size on the socket
    wasn't enough; we need to limit readBody to the available amount of bytes
    in the QNetworkReply device, so that it waits until the app has read data.
    And then to trigger more reading when that device has room again. And to
    flush all that is left when the socket disconnects.
    
    Unit-test included, which tests 4 cases (http/https, with/without buffer size)
    and checks the send/receive rates to check that the actual transfer rate
    was limited, not just the speed of emission of data to the application.
                                                                                  
    Task-number: QTBUG-6276 </proposal>
    <summary>QNAM HTTP Code: support QNetworkReply::setReadBufferSize() when downloading.</summary>
    <updated-at type="datetime">2009-12-15T09:55:36Z</updated-at>
    <status>Reviewing</status>
    <username>~dfaure</username>
    <source_repository>
      <name>dfaures-clone</name>
      <branch>qnam_throttling</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-10T15:43:09Z</created-at>
    <ending-commit>b5e59e4bd863e9824ba6b68159db80639c3af1b4</ending-commit>
    <id type="integer">2366</id>
    <proposal>    This is done on all parts that are exported (objdump) and
    documented (qdoc3). There are more places where this should be
    changed as the class should always be stated explicitly.

Example:
-    Direction direction() const;
-    void setDirection(Direction direction);
+    QAbstractAnimation::Direction direction() const;
+    void setDirection(QAbstractAnimation::Direction direction);
</proposal>
    <summary>Changing enum types in functions from "enum" to contain "class::enum"</summary>
    <updated-at type="datetime">2009-12-10T15:43:26Z</updated-at>
    <status>New</status>
    <username>~ulf</username>
    <source_repository>
      <name>bc-qt</name>
      <branch>subclass_fix</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-09T13:27:20Z</created-at>
    <ending-commit>dbf58506e7e2f9d93c4136e82cb8d9048cd04a59</ending-commit>
    <id type="integer">2357</id>
    <proposal>When popup was opened SetPointerCapture(..) was always cancelled, but when it was closed it was not given back to the parent dialog. When modal dialog was open and user clicked outside the dialog there was no check if the dialog is modal. Which caused wrong behaviour. Check for the modal dialog was also added.</proposal>
    <summary>Dialog loses its modality if user opens a popup within the dialog</summary>
    <updated-at type="datetime">2010-02-06T14:06:59Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~joonas</username>
    <source_repository>
      <name>cask-upstream</name>
      <branch>QTBUG-4810</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-09T09:55:10Z</created-at>
    <ending-commit>ba93d25091d3596c9b6a41def1ed6048ad13ab55</ending-commit>
    <id type="integer">2354</id>
    <proposal>This is intended to be used in interoperability scenarios, where the time values are read from a file or network connection that provides the FILETIME structure values (two 32-bit values), and this allows us to handle those values as a convenient QDateTime object.


</proposal>
    <summary>Additional constructor to support conversion from Windows FILETIME</summary>
    <updated-at type="datetime">2009-12-09T09:55:14Z</updated-at>
    <status>New</status>
    <username>~bradh</username>
    <source_repository>
      <name>bradh-clone</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-07T22:32:02Z</created-at>
    <ending-commit>0355d65c1f584d99ac2e916b262268e30409502e</ending-commit>
    <id type="integer">2341</id>
    <proposal>Remove obsolete function set_winapp_name() and non-Unicode appFileName &amp; theAppName, which are not used anywhere.

Make sure that qWinAppInst(), qWinAppPrevInst(), qWinAppCmdShow() always return correct values. Do not rely on the values passed by WinMain.</proposal>
    <summary>Cleanup set_winapp_name()</summary>
    <updated-at type="datetime">2009-12-07T22:32:17Z</updated-at>
    <status>New</status>
    <username>~miniak</username>
    <source_repository>
      <name>miniak</name>
      <branch>set_winapp_name</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-07T22:29:05Z</created-at>
    <ending-commit>356648d6863ff23d4d118430dc13f54ae856df5b</ending-commit>
    <id type="integer">2340</id>
    <proposal>most important of them:

*) Unicode character property 'General_Category=Cn' was erroneously ignored causing two related bugs in QChar::category() and QChar::isPrint().

As described in tr44, 4.2.8 Default Values 
General_Category character property should be setted to Cn 
for all codepoints omitted in UCD. 
Instead they was mapped to QChar::NoCategory (=0) which have no equivalent 
in Unicode specs and as result was ignored in filtering by binary mask.

*) reduce count of the unique unicode properties by fixing specialCaseMap entries uniqueness

+) some cleanups and minor improvements

http://qt.gitorious.org/qt/qt/merge_requests/1340 is hardly depends on this MR</proposal>
    <summary>major fixes for qunicodetables data and data generator</summary>
    <updated-at type="datetime">2010-01-10T13:30:04Z</updated-at>
    <status>New</status>
    <username>~ritt.k</username>
    <source_repository>
      <name>ritts_qt</name>
      <branch>fixes/qunicodetables-what_ive_always_wanted</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-07T12:59:11Z</created-at>
    <ending-commit>786734e4dbfebd9dac70fe913c8b94f3a7d2b939</ending-commit>
    <id type="integer">2336</id>
    <proposal></proposal>
    <summary>add support of high unicodes to QUrl</summary>
    <updated-at type="datetime">2010-01-09T20:23:02Z</updated-at>
    <status>Reviewing</status>
    <username>~ritt.k</username>
    <source_repository>
      <name>ritts_qt</name>
      <branch>improvements/qurl/support_high_unicodes</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-03T13:45:11Z</created-at>
    <ending-commit>9a7c58380028d60253608bc31d6eccdf4e8bf483</ending-commit>
    <id type="integer">2292</id>
    <proposal>Q3ListViewItem::sortChildItems() exits early if there is only 1 child
item and leaves grandchildren unsorted.

</proposal>
    <summary>Fixing bug with q3listview sorting</summary>
    <updated-at type="datetime">2010-02-09T13:37:29Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~akaraivanov</username>
    <source_repository>
      <name>mi-clone</name>
      <branch>4.6-q3listview-sorting-bug</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-02T15:18:43Z</created-at>
    <ending-commit>4eba5ca5a16160136010f291b58919d5e786ecb8</ending-commit>
    <id type="integer">2282</id>
    <proposal>Allow building Qt without QtGui

This patch adds a -no-gui option to configure, allowing to build just QtCore and other non-GUI dependant parts (QtNetwork, QtXml, ...).

This allows building non-GUI parts on machines that don't have X11
headers installed (e.g. headless servers) and building non-GUI parts
on platforms where the GUI part can't be compiled yet.
</proposal>
    <summary>Make building QtGui optional</summary>
    <updated-at type="datetime">2009-12-18T14:21:50Z</updated-at>
    <status>Reviewing</status>
    <username>~blankpage</username>
    <source_repository>
      <name>blankpage-patches</name>
      <branch>optionalize-qtgui</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-02T09:55:42Z</created-at>
    <ending-commit>0f2540791c89eb53ccd142f40bc465919b80e607</ending-commit>
    <id type="integer">2279</id>
    <proposal>Fixes: QTBUG-4670 Keypad navi focus rect is not visible around QFileDialog's itemview.
    
Qfiledialog was modified to add focusframe margins to layout depending on viewmode and if symbian os and keypad navigation are defined.

http://bugreports.qt.nokia.com/browse/QTBUG-4670

Task number: QTBUG-4670</proposal>
    <summary>Navi focus rect made visible around qfiledialog itemview</summary>
    <updated-at type="datetime">2009-12-02T09:55:49Z</updated-at>
    <status>New</status>
    <username>~manolvi</username>
    <source_repository>
      <name>cask-upstream</name>
      <branch>QTBUG-4670</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-12-01T07:38:10Z</created-at>
    <ending-commit>3bcc209d4f4918dcb914f0ec9e27378137544f66</ending-commit>
    <id type="integer">2270</id>
    <proposal>QFocusFrame::eventFilter resize event was modified so that it will call show()
which will make focusframe visible when changing orientation</proposal>
    <summary>Keypad navi focus disappears when orientation change occurs</summary>
    <updated-at type="datetime">2009-12-01T07:38:11Z</updated-at>
    <status>New</status>
    <username>~joonas</username>
    <source_repository>
      <name>cask-upstream</name>
      <branch>QTBUG-4811</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-30T13:22:56Z</created-at>
    <ending-commit>51a847df916a424c349782ad9bd3d7f990f53590</ending-commit>
    <id type="integer">2263</id>
    <proposal>Under X11 release events for Qt::XButton1 and Qt::XButton2 mouse buttons are not detected and this patch adds support for them.</proposal>
    <summary>Release events for Qt::XButton1 and Qt::XButton2.</summary>
    <updated-at type="datetime">2009-12-15T10:36:54Z</updated-at>
    <status>Reviewing</status>
    <username>~mirko</username>
    <source_repository>
      <name>qt-mirko</name>
      <branch>xbuttonsupp</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-27T14:56:05Z</created-at>
    <ending-commit>12638561109a6da9d63817e95b8b4c72cff69182</ending-commit>
    <id type="integer">2236</id>
    <proposal>We shouldn't cache not only executables and symlinks but the icons files too. Otherwise first .ico file will be cached for this extension and all the other icons will have the same appearance.</proposal>
    <summary>Exclude .ico files from QFileIconProvider's cache</summary>
    <updated-at type="datetime">2009-11-27T14:56:07Z</updated-at>
    <status>New</status>
    <username>~uriel</username>
    <source_repository>
      <name>uriels-qt</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-25T10:10:27Z</created-at>
    <ending-commit>ccd0a43a41510934a0f92fe8f0282098846b4b22</ending-commit>
    <id type="integer">2207</id>
    <proposal>Minimizing widget left statuspane and softkeys visible, focus was also
left to the minimized widget. In this fix widget is sent to background
cause Symbina does not have a feature for minimizing window.
Reviewed by Markku Luukkainen</proposal>
    <summary>Minimized widgets have statusbar and softkey.</summary>
    <updated-at type="datetime">2009-12-14T10:03:02Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~joonas</username>
    <source_repository>
      <name>cask-upstream</name>
      <branch>QTBUG-5047</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-23T22:44:26Z</created-at>
    <ending-commit>71ff7d306296adf75265d131b9330c0e1da0efe7</ending-commit>
    <id type="integer">2193</id>
    <proposal>Hello :)

This patch implements PCH support for CONFIG=universal on OSX.

It is implemented as described below:

  - A new method for the Unix Makefile generator is added: buildArchitectures(). It returns a list of architectures the target is to be built for. In case we're on a system without support for universal binaries, or we're on a non-universal build, it returns a list consisting of an empty QString.

  - All places where the Makefile generator outputs PCH-related code, we loop through all our build architectures. If we detect that we're being built for more than one architecture (i.e. a universal build), we append the architecture name to the precompiled header file. E.g.:

    Mumble.gch becomes Mumble-i386.gch

 - A new variable is added to the mkspecs of build configurations that wish to support this: QMAKE_CFLAGS_USE_PRECOMPILE_UNIVERSAL. It is universal variant of QMAKE_CFLAGS_USE_PRECOMPILE. 

 - The aforementioned variable is implemented for GCC-4.2 using -Xarch. This allows us to keep the "Apple"-way of building universal binaries; that is, using a single 'gcc -arch i386 -arch ppc -c foo.o foo.c [...]' commandline to compile to a 'universal' object that includes both a i386 section, and a ppc section.

- The universal variant of USE_PRECOMPILE using Xarch looks like this: -Xarch_i386 -include -Xarch_i386 ${QMAKE_PCH_OUTPUT_BASE}-i386 [...], and as such, it depends on the (hard-coded) practice of simply appending "-$arch" to the 'regular' PCH filename.

Since we now support universal binary PCH, we remove the checks that deny us this ability from configure.

Because we use -Xarch, we require Apple GCC 4.2 (or greater). The old conditional for disabling PCH for universal builds has been changed to only occur when we're on a compiler that doesn't support -Xarch.

This patch has not been tested on llvm-gcc (but I don't see why it wouldn't work...), nor ICC (which maybe is a little more tricky!)

Feedback very much welcomed.</proposal>
    <summary>Allow universal binary builds on Mac OS X to use precompiled headers</summary>
    <updated-at type="datetime">2009-11-23T22:44:27Z</updated-at>
    <status>New</status>
    <username>~mkrautz</username>
    <source_repository>
      <name>mkrautzs-qt-clone</name>
      <branch>macx-universal-pch</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-21T18:45:45Z</created-at>
    <ending-commit>ff912c80db85aea867798c7fc6c7296584945bae</ending-commit>
    <id type="integer">2175</id>
    <proposal>I may have submitted this once before, but I blew away my repo and recreated it, so I think it got lost...

There are several places that blindly #define _WIN32_WINNT without checking to see if it is already defined.  This patch appropriately handles the checking (and redefining in the case of qwindowsxpstyle.cpp).</proposal>
    <summary>Make sure _WIN32_WINNT is not redefined</summary>
    <updated-at type="datetime">2009-11-21T18:45:47Z</updated-at>
    <status>New</status>
    <username>~gordonschumacher</username>
    <source_repository>
      <name>qt-rebit</name>
      <branch>win32-winnt-redef</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-17T21:57:19Z</created-at>
    <ending-commit>e13e837874f7f16ee5e67927d456f701b5383cee</ending-commit>
    <id type="integer">2126</id>
    <proposal>On Unix-like systems Qt can be linked to zlib version already available in the system. On Windows configuration program accepts the "-system-zlib" option, but project files don't contain appropriate code.

This patch updates Qt project files to allow usage of system zlib on Windows platforms. zlib import library is assumed to be named "zdll.lib" for MSVC compiler (this name is used in official zlib binary and source packages) or "libz.dll.a" for MinGW.</proposal>
    <summary>Added the ability to use system zlib on Windows</summary>
    <updated-at type="datetime">2009-12-06T00:43:04Z</updated-at>
    <status>New</status>
    <username>~dinosaur</username>
    <source_repository>
      <name>dinosaurs-clone</name>
      <branch>win32-system-zlib</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-16T13:52:18Z</created-at>
    <ending-commit>ead1c00d65feed46106efc58bd50d197a191bef4</ending-commit>
    <id type="integer">2105</id>
    <proposal>On Windows with a shared library project, if you currently do a "make install" only the .lib and .dll files get installed to the target.path. It is very useful to also have the debugging symbols in the .pdb file available when the library is built with a debug configuration.

This patch enables this.</proposal>
    <summary>Make qmake install debugging symbols on windows</summary>
    <updated-at type="datetime">2010-01-21T21:14:05Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~seanharmer</username>
    <source_repository>
      <name>seanharmers-qt</name>
      <branch>qmake-install-pdb</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-08T14:37:13Z</created-at>
    <ending-commit>36e6a8811d32a6da502da15e20b1efa995ba0a52</ending-commit>
    <id type="integer">2034</id>
    <proposal>On a listview in the icon mode with tab navigation enabled, the tab moved focus down (even though the flow() is LeftToRight), and stopped after hitting the bottommost item in that vertical.

Methinks it should just move to the next item, wherever that's laid out. Hence this change.</proposal>
    <summary>more natural tabbing in QListView</summary>
    <updated-at type="datetime">2009-11-09T12:38:42Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~roop</username>
    <source_repository>
      <name>qt-roop</name>
      <branch>iconview_changes2</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-11-05T02:27:55Z</created-at>
    <ending-commit>4a4c4b02202cca6cb35a114408f9f97786e8abae</ending-commit>
    <id type="integer">2013</id>
    <proposal>This change moves the call to qwsEventFilter to be a bit earlier in qwsProcessEvent, so that blocked mouse events truly have no affect on the app (without this change, a blocked mouse event still would cause the app to change the "last known" button state in the mouse_state variable)
</proposal>
    <summary>call qwsEventFilter sooner</summary>
    <updated-at type="datetime">2010-02-06T14:45:41Z</updated-at>
    <status>Revise and resubmit</status>
    <username>~nordellj</username>
    <source_repository>
      <name>qt-chloride</name>
      <branch>qws-event-filter-tweak</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.5</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-10-31T03:12:43Z</created-at>
    <ending-commit>3ea5afd702acd770477a3c7f444a461947091d1e</ending-commit>
    <id type="integer">1972</id>
    <proposal>if use code like this, the isValid method will never be updated if server closes the connection.

	QLocalSocket l_localSocket;
	QByteArray l_ba;

	while(l_localSocket.isValid())
	{
		if (l_localSocket.waitForReadyRead(100) == true)
		{
			l_ba.append(l_localSocket.readAll());
		}
	....
	....
	....
	....
	}
</proposal>
    <summary>Windows QLocalSocket::isValid() 'is Pipe closed' check added</summary>
    <updated-at type="datetime">2009-11-16T14:02:44Z</updated-at>
    <status>Reviewing</status>
    <username>~bokic</username>
    <source_repository>
      <name>bokics-clone</name>
      <branch>master</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-10-30T20:02:58Z</created-at>
    <ending-commit>a2cdf90e150866fa307ea2ddda9e4c04c982f647</ending-commit>
    <id type="integer">1969</id>
    <proposal>This adds a loadHints property to the QScriptEngine class that works the same way as the property of the same name on QPluginLoader and allows to provide hints used when QScriptEngine::importExtension() loads a QScript plugin.

Why this is necessary:

Under the gcc abi, exception handling and RTTI only work as intended in explicitly loaded shared libraries on unix platforms if RTLD_GLOBAL is passed to dlopen when the library is loaded (cf http://gcc.gnu.org/faq.html#dso ).

QLibrary and QPluginLoader can be set to pass this flag to dlopen through their loadHints property, but QScriptEngine lacked a way to specify this when loading QtScript plugins, which is problematic for plugins that internally use exceptions or RTTI.

This is a resubmit of merge request 1856 with the requested changes for the master branch.</proposal>
    <summary>Allow to provide QLibrary::loadHints for QScriptEngine::importExtension() - resubmit for master branch</summary>
    <updated-at type="datetime">2009-10-30T20:03:00Z</updated-at>
    <status>New</status>
    <username>~Zlodo</username>
    <source_repository>
      <name>zlodos-qt-clone</name>
      <branch>extensionloadhints</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>master</branch>
    </target_repository>
  </merge-request>
  <merge-request>
    <created-at type="datetime">2009-10-30T07:46:15Z</created-at>
    <ending-commit>1875fa979bf498c129f975ed1ba1375865307740</ending-commit>
    <id type="integer">1962</id>
    <proposal>In SVG, polyline should be used more for polylines than a straight line, since line is available in SVG then use this instead for drawing single lines.</proposal>
    <summary>Resubmit of #1869 - When painting a line in QSvgGenerator, then use &lt;line&gt; tag instead of &lt;polyline&gt; </summary>
    <updated-at type="datetime">2009-10-30T07:46:17Z</updated-at>
    <status>New</status>
    <username>~andyshaw</username>
    <source_repository>
      <name>andyshaws-clone</name>
      <branch>svgline</branch>
    </source_repository>
    <target_repository>
      <name>qt</name>
      <branch>4.6</branch>
    </target_repository>
  </merge-request>
</merge-requests>
